One of the most impressive changes in economics and decision sciences is the emergence and rapid growth of so-called “behavioral” economics and neuroeconomics. These fields raise several methodological issues, some of which are at the forefront of current debate. Among those issues, the most prominent is: what is the epistemic relevance of non-behavioral or “cognitive” data, i.e., data which bear on cognitive processes and states involved in decision making? Faruk Gul and Wolfgang Pesendorfer (2005/2008) have vigorously criticized the idea that these data could be relevant for economics and decision sciences. Their criticisms became the focal point of a very active methodological literature. In this paper, we reconstruct and discuss Gul and Pesendorfer’s views and arguments. Although we are not convinced by some of them, we believe they suggest a genuine issue: the “missing links.”